Media outlets, such as broadcast television and radio stations affiliated with, for example, NBC, ABC, CBS, FOX, CNN, ESPN, and so on, generally deliver media in the form of media streams. Such media streams link together numerous media objects having a variety of different lengths and conveying information on a variety of different topics. For example, a CNN station (radio or television) might have a 30 minute program (i.e., media stream) that includes various news clips, followed by various advertisement clips, followed by various weather-related news clips, followed again by various advertisement clips, followed by various sports-related news clips, and so on. Similarly formatted media streams might repeat on a CNN station every 30 minutes.
The information available in such media streams from different media outlets is generally comprehensive. That is, given enough time, particular media objects within media streams, such as songs or certain news clips, will likely begin to repeat within the media streams. For example, an important news story on a CNN station may begin as a breaking news clip. The breaking story might then be repeated in various forms throughout the day during each 30 minute news segment broadcast on the CNN station. Thus, when an important news story initially emerges, the breaking news clip covering the story might initially play in its entirety while subsequent coverage of the same story may only include repeated, but smaller versions of the original breaking news clip.
Various other scenarios can be described in which media objects in media streams exhibit repeat patterns within the media streams. For example, some news stories start out being of minor importance, but then continue to grow into very important stories. Thus, initial coverage of such a story might include a short news clip, while subsequent coverage may include longer and more in-depth news clips containing additional information about various aspects of the story. As the story begins to lose importance, the length and breadth of the news clips covering the story diminish again. However, throughout the coverage of the story, it is likely that the various news clips contain repeated portions of prior news clips. Such repeated portions can be recognized as repeating media objects within a media stream, as is generally described in the related application, “A System And Method For Providing User Control Over Repeating Objects Embedded In A Stream”, as noted above.
In another similar example, a music media outlet (e.g., a radio station) may broadcast a debut song from a new artist on a relatively infrequent basis. However, the song may quickly become a popular hit, and the media outlet might dramatically increase the frequency with which it broadcasts the new song. As the popularity of the new song wears off, the frequency with which it is broadcast would also decrease. Again, the repeating song within the media outlet's media stream can be recognized as a repeating media object.
In addition to being comprehensive, however, the information available in media streams from various media outlets is diffuse. As mentioned above, the types of information in such media streams can cover a wide variety of different topics. The wide variety of information available in a media stream presents a problem when trying to find a particular media object within the stream. Furthermore, trying to discover additional information about a particular media object within a media stream presents an even greater challenge. Current methods for identifying media objects within a media stream include creating a database of known objects that can be compared (e.g., using audio fingerprinting) against sampled portions of the media stream. However, unless a database of known objects is created or already available for comparison, such methods are not possible. The alternative is to locate a desired media object from within a media stream by “manually” sifting through a vast amount of data, which is impractical.
The related application, “A System And Method For Providing User Control Over Repeating Objects Embedded In A Stream”, mentioned above, includes a discussion of identifying and extracting or segmenting repeating media objects from a media stream. A need exists, however, for a way to infer information about such media streams in general, and about the media objects within such media streams more specifically.